AI is described as a major unlock for lean, IC-heavy PM orgs (see "We Regret That Product Management Exists" (Whatnot's PM-Scarcity Principle), IC-First Product Leadership (Keep Top PMs Doing the Work)): PMs can now self-serve work that used to require another specialist.
Downside: data scientists report now spending time reviewing "half-assed" AI-driven analysis produced by non-specialist PMs, rather than doing original analytical work themselves — the self-service gain for PMs creates a review burden elsewhere.
Related: AI-Driven Bifurcation of Product Manager Roles, AI as Amplifier of Existing Discovery Culture, Two Distinct 'Data' Roles (Analyst vs. ML Engineer).
Rather than asking an engineer to scope a feature or explain how a system works, a PM can query an AI assistant (e.g., Claude Code) directly against the codebase for level-of-effort estimates and system logic — reserving the human scoping conversation for after that first pass. See also Hex Threads (Self-Serve PM Data Tool) for the data-side equivalent.
The self-service gain for PMs has a cost on the other side. As PMs do more of their own data pulls and analysis with AI tools and talk to data scientists less for routine questions, data scientists increasingly spend their time auditing lower-quality AI-assisted analysis that PMs produced, rather than doing original analysis themselves. The net effect isn't simply "less data science work needed" — it's a shift in what data science work is, toward review and correction.
Related: Two Distinct 'Data' Roles (Analyst vs. ML Engineer).
At LinkedIn, under the LinkedIn's Full-Stack Builder Model, designers and PMs have started picking up bugs directly from Jira and submitting the pull requests themselves — cited by CPO Tomer Cohen as one of the earliest concrete wins of the transformation (see Experimentation Velocity Formula: (Volume × Quality) ÷ Time-to-Launch).
Beyond PMs using AI for data and debugging, the underlying claim is that within 5–10 years tools like Figma AI, GitHub Copilot, and Cursor will erode the sharp boundaries between product, engineering, and design roles altogether — not just give PMs new self-service capabilities but make the traditional handoff between 'who specifies' and 'who builds' increasingly optional. Practical implication: PMs and designers should use these tools to build their own prototypes or working code rather than treating implementation as strictly an engineer's job. See AI-Driven Bifurcation of Product Manager Roles and Storming vs. Norming: AI's Upheaval in Product Roles (Elizabeth Stone) for related but distinct claims about how AI reshapes the PM role itself.